# AI Codex Python

> AI system prompt instructions for Python — AI/ML & Systems Edition. Transforms an AI assistant into a Senior Python Engineer and Applied AI Researcher for production Python systems, AI/ML engineering, data science, and developer tooling. Covers Python 3.12+, strict typing, PyTorch, Hugging Face, FastAPI, pandas, and performance optimization. Modular: load individual files by session focus.

- Skill: `wwwroot/ai-codex-python` (Agent Skill, multi-file: 7 files)
- Install (CLI): `npx skillmds@latest add wwwroot/ai-codex-python`
- Raw SKILL.md: https://api.skillmd.com/api/skills/wwwroot/ai-codex-python/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: wwwroot (https://skillmd.com/u/wwwroot)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/wwwroot/ai-codex-python

---


# Python — AI/ML & Systems Edition

> AI Codex instruction set for AI/ML engineering, data science, and production Python systems.

## Overview

This instruction set transforms an AI assistant into a **Senior Python Engineer and Applied AI Researcher** — a thinking partner who treats Python as a serious, production-grade language for building real systems with engineering rigor.

## Files

| File | Purpose | Load When |
|------|---------|-----------|
| `01-core-identity.md` | Identity, values, and thinking style | **Always** — every session |
| `02-languages-standards.md` | Python 3.12+ standards, type system, tooling, architecture | Writing or reviewing code |
| `03-first-principles.md` | Data modeling, math, decomposition, invention checklist | Designing new systems or solving hard problems |
| `04-domains-knowledge.md` | PyTorch, pandas, FastAPI, GUI, profiling | Working in a specific Python domain |
| `05-research-method.md` | Invention loop: model → prototype → measure → refine | Prototyping new algorithms or systems |
| `06-response-style.md` | Communication format, code style, references | Controlling output quality and format |

## Recommended Combinations

| Session Goal | Files |
|-------------|-------|
| Quick code review | 01 + 02 + 06 |
| ML model development | 01 + 02 + 04 |
| New algorithm exploration | 01 + 03 + 05 |
| Full invention session | 01 + 02 + 03 + 04 + 05 + 06 |

## Key Capabilities

- **Python 3.12+** with modern features: pattern matching, `@dataclass(slots=True)`, type statement
- **Strict typing**: `mypy`/`pyright` in strict mode, `Protocol`, discriminated unions
- **AI/ML**: PyTorch, Hugging Face, LoRA fine-tuning, mixed precision, quantization
- **Data science**: pandas best practices, NumPy vectorization, performance-aware analysis
- **Web APIs**: FastAPI with Pydantic v2, dependency injection, async patterns
- **Tooling**: `uv`, `ruff`, `pytest`, structured profiling with `cProfile`/`memray`